CN115327373A - Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium - Google Patents
Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium Download PDFInfo
- Publication number
- CN115327373A CN115327373A CN202210416019.9A CN202210416019A CN115327373A CN 115327373 A CN115327373 A CN 115327373A CN 202210416019 A CN202210416019 A CN 202210416019A CN 115327373 A CN115327373 A CN 115327373A
- Authority
- CN
- China
- Prior art keywords
- fault
- neural network
- motor
- instrument
- training sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/34—Testing dynamo-electric machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Control Of Electric Motors In General (AREA)
- External Artificial Organs (AREA)
Abstract
The application relates to a hemodialysis equipment fault diagnosis method and a storage medium based on a BP neural network, relating to the technical field of motor fault detection and comprising the following steps: acquiring a current spectrogram of an instrument motor to identify fault symptoms of the instrument motor; obtaining an amplitude corresponding to a fault characteristic frequency component of a fault symptom and an amplitude of a fundamental component, and taking a ratio of the amplitude corresponding to the fault characteristic frequency component and the amplitude of the fundamental component as an input layer neuron of the BP neural network; acquiring a fault mode of an instrument motor to serve as an output layer neuron of a BP neural network; acquiring a standard training sample set and a test sample set, and obtaining a standard training sample test result corresponding to the standard training sample set and a test sample test result corresponding to the test sample set based on a BP neural network; and comparing the test result of the training sample with the test sample result to obtain the fault mode of the motor of the instrument. This application has the effect that improves the detection accuracy to taking motor instrument trouble.
Description
Technical Field
The application relates to the field of motor fault detection, in particular to a hemodialysis equipment fault diagnosis method based on a BP neural network and a storage medium.
Background
In the related art, in terms of medical instruments, a servo motor is arranged in a plurality of instruments, for example, a hemodialysis machine is a treatment instrument which is most widely applied in blood purification treatment, and is a relatively complex electromechanical integrated device which consists of a supply monitoring device and an extracorporeal circulation monitoring device. The operating principle is that the dialysate is used to carry out solute dispersion, osmosis and ultrafiltration with the patient blood led out by the equipment through a hemodialyzer; the blood of the patient after the action returns to the body of the patient through the equipment pipeline, and the liquid after dialysis is taken as waste liquid and is discharged by the dialysate supply system; the continuous circulation is carried out, thereby achieving the purpose of treatment and completing the whole dialysis process.
The main components of a hemodialysis apparatus are motors, such as blood pumps, heparin pumps, dialysate pumps, etc., which are the power to drive the extracorporeal circulation of dialysate and blood. The hemodialysis machine is different from other common electronic instruments, and the machine is stopped when a pump fails in the dialysis process, so that the treatment work and even the safety of patients are influenced. Therefore, it is very important to ensure the normal operation of the machine, especially the various motors as main components, and therefore, how to improve the accuracy of detecting the fault of the motor-equipped apparatus is urgently needed.
Disclosure of Invention
In order to improve the detection accuracy of the faults of the belt motor instrument, the application provides a hemodialysis equipment fault diagnosis method based on a BP neural network and a storage medium.
In a first aspect, the hemodialysis device fault diagnosis method based on the BP neural network provided by the application adopts the following technical scheme:
the hemodialysis equipment fault diagnosis method based on the BP neural network is characterized by comprising the following steps of:
acquiring a current spectrogram of an instrument motor to identify fault symptoms of the instrument motor;
obtaining an amplitude value corresponding to a fault characteristic frequency component of a fault symptom and an amplitude value of a fundamental component, and taking a ratio of the amplitude value corresponding to the fault characteristic frequency component and the amplitude value of the fundamental component as an input layer neuron of the BP neural network;
acquiring a fault mode of an instrument motor to serve as an output layer neuron of a BP neural network, wherein the fault mode is used for representing the fault type of the instrument motor;
acquiring a standard training sample set and a test sample set, and obtaining a standard training sample test result corresponding to the standard training sample set and a test sample test result corresponding to the test sample set based on a BP neural network;
and comparing the test result of the training sample with the test sample result to obtain the fault mode of the motor of the instrument.
By adopting the technical scheme, the motor at the initial stage of the fault often generates different forms of fault symptoms such as mechanical abnormal vibration, abnormal current signal change and the like, the fault symptoms are preliminarily identified through the current spectrogram of the motor of the instrument, and the fault of the motor is accurately detected and diagnosed through the BP neural network, so that the fault mode of the motor is rapidly and accurately diagnosed, the time spent on motor maintenance is reduced, and the accuracy of fault detection of the instrument with the motor is greatly improved.
Preferably, the step of obtaining a current spectrogram of the instrument motor to identify a fault symptom of the instrument motor includes:
comparing a current frequency component image of a stator winding obtained by detecting an asymmetric motor rotor in an instrument with a preset variable frequency band image to determine a rotor winding fault symptom, wherein the preset variable frequency band is a variable frequency band with the size of 2sf and appears on two sides of a fundamental wave;
comparing a current harmonic image obtained by detecting a relatively static stator in an instrument with a preset current harmonic image to determine an air gap eccentricity fault symptom, wherein the preset current harmonic is a fundamental wave and variable frequency bands appear on two sides of a main tooth harmonic;
and obtaining a test harmonic image based on the stator current and the air gap magnetic field detection, and comparing the test harmonic image with a preset harmonic image to determine the stator winding fault symptom, wherein the preset harmonic image is the nth harmonic and the mth harmonic increase of the current.
By adopting the technical scheme, the corresponding current frequency component image, the corresponding current harmonic image and the corresponding test harmonic image are obtained from the current spectrogram of the instrument motor, and the current frequency component image is compared with the preset variable frequency band image, so that the rotor winding fault symptom is determined; comparing the current harmonic image with a preset current harmonic image to determine an air gap eccentricity fault symptom; and comparing the test harmonic image with a preset harmonic image to determine the fault symptom of the rotor winding, and facilitating the preliminary judgment of the fault part of the motor by a detector or a worker by observing the current spectrogram of the motor of the instrument.
Preferably, the step of obtaining the amplitude corresponding to the fault characteristic frequency component of the fault symptom and the amplitude of the fundamental component, and using a ratio between the amplitude corresponding to the fault characteristic frequency component and the amplitude of the fundamental component as an input layer neuron of the BP neural network includes:
acquiring amplitudes corresponding to frequency components f, (1 + 2S) f, (1-2S) f, [1+ (1-S)/p ] f, [1- (1-S)/p ] f,3f and 5f and amplitudes of fundamental wave components in a current spectrogram of an instrument motor;
calculating the ratio between the amplitude corresponding to the fault characteristic frequency component and the amplitude of the fundamental component based on a ratio algorithm;
and taking the ratio as an input layer neuron of the BP neural network.
Preferably, the step of acquiring the failure mode of the instrument motor as an output layer neuron of the BP neural network includes:
acquiring various failure modes of an instrument motor;
establishing a fault matrix based on the fault pattern, wherein elements of a row vector of the fault matrix correspond to the fault pattern;
and establishing an incidence relation between the fault matrix and output layer neurons of the BP neural network, wherein [1,0,0], [0,1,0], [0,0,1] are used as fault modes of corresponding output layer neurons, zero-value elements represent no faults, and non-zero-value elements represent corresponding fault modes.
By adopting the technical scheme, the fault mode is associated with the fault matrix, the [1,0,0], [0,1,0], [0,0,1] is taken as the fault mode of the corresponding neuron of the output layer, the zero value element represents no fault, and the non-zero value element represents the corresponding fault mode, so that the fault mode can be conveniently judged by a worker and a computer.
Preferably, after determining the input layer neurons and the output layer neurons of the BP neural network, the method further comprises:
determining the number of nodes of an input layer and an output layer based on the dimension required by the input vector and the output vector of the BP neural network;
determining the number of neurons of the hidden layer based on a relevant empirical algorithm;
transfer functions between layers of the BP neural network are determined.
By adopting the technical scheme, the construction of the BP neural network is further perfected.
Preferably, the relevant empirical algorithm comprises
Wherein k is the number of samples, M is the number of neurons of a hidden layer of the BP neural network, and N is the number of neurons of an input layer;
and for all i > M, C is true M ＝0；
Wherein m is the neuron number of the input layer, n is the neuron number of the output layer, and a is [0,10 ]]Any constant within the interval.
By adopting the technical scheme, the performance of the whole BP neural network can be basically determined by the number of the neurons of the hidden layer, generally, the more the number of the neurons of the hidden layer is, the better the network performance is, however, if the number of the neurons of the hidden layer is too much, the training time of the whole BP neural network is too long, and the number of the neurons is determined by the formula, so that the performance of the BP neural network is improved conveniently.
Preferably, the step of determining a transfer function between layers of the BP neural network comprises:
adopting a sigmoid function as a transfer function of a hidden layer;
and taking a Log-sigmoid function as an output layer transfer function based on the output state of the output layer neuron.
By adopting the technical scheme, the range of the function value can be compressed to [0,1], and data can be compressed without changing the amplitude.
Preferably, the step of obtaining the standard training sample set and the test sample set, and obtaining the standard training sample test result corresponding to the standard training sample set and the test sample test result corresponding to the test sample set based on the BP neural network, includes:
acquiring a standard training sample set and a test sample set;
obtaining a standard training sample test result of the standard training sample set based on a BP neural network;
obtaining a test sample test result of the standard training sample set based on a BP neural network;
comparing the test result of the standard training sample with the test result of the test sample to obtain a comparison result;
if all the comparison results are zero values, the instrument with the motor is normal and has no fault;
if the comparison result has a non-zero value, the motor instrument is indicated to have a fault.
By adopting the technical scheme, the standard training sample test result and the test sample test result are compared to obtain a comparison result, and whether the motor instrument has a fault is judged by adopting a mode of judging whether the zero value exists, so that the working state of the motor instrument can be judged by a worker conveniently.
Preferably, the standard training sample set comprises a rotor winding fault sample, an air gap eccentricity fault sample, a stator winding fault sample and a normal standard sample;
the set of test samples is a set of samples from actual measurements.
In a second aspect, the present application provides a computer-readable storage medium, which adopts the following technical solutions:
the storage medium stores at least one instruction, at least one program, a set of codes, or a set of instructions that is loaded and executed by the processor to implement any one of the BP neural network based hemodialysis device fault diagnosis methods described above.
In summary, the present application includes at least one of the following beneficial technical effects:
1. the motor at the initial stage of the fault often generates different forms of fault symptoms such as mechanical abnormal vibration, abnormal current signal change and the like, the fault symptoms are preliminarily identified through a current spectrogram of the motor instrument, and the fault of the motor is accurately detected and diagnosed through a BP neural network, so that the fault mode of the motor is rapidly and accurately diagnosed, the time spent on motor maintenance is reduced, and the fault detection accuracy of the motor instrument is greatly improved;
2. the standard training sample test result and the test sample test result are compared to obtain a comparison result, and whether the motor instrument has a fault is judged by judging whether a zero value exists or not, so that the working state of the motor instrument can be judged more conveniently by a worker.
Drawings
Fig. 1 is a schematic flow chart illustrating steps of a hemodialysis device fault diagnosis method based on a BP neural network according to an embodiment of the present application;
FIG. 2 is a schematic flow chart of the substeps of step S1;
FIG. 3 is a schematic flow chart of the substep of step S2;
FIG. 4 is a flow chart illustrating the substep of step S3;
FIG. 5 is a flowchart illustrating the substep of step S4;
FIG. 6 is a diagram of the tansig function;
FIG. 7 is a logsig function.
Detailed Description
The present application is described in further detail below with reference to figures 1-5.
The present application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
In the following description, for purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the inventive concepts. Some of the figures in the present disclosure show structures and devices in block diagram form as part of this specification to avoid obscuring the disclosed principles. In the interest of clarity, not all features of an actual implementation are described in this specification. Moreover, the language used in the present disclosure has been principally selected for readability and instructional purposes, and may not have been selected to delineate or circumscribe the inventive subject matter, resort to the claims being necessary to determine such inventive subject matter. Reference in the present disclosure to "one implementation" or "an implementation" means that a particular feature, structure, or characteristic described in connection with the implementation is included in at least one implementation, and references to "one implementation" or "an implementation" are not to be understood as necessarily all referring to the same implementation.
The embodiment of the application discloses a hemodialysis equipment fault diagnosis method based on a BP neural network.
Referring to fig. 1, the hemodialysis device fault diagnosis method based on the BP neural network includes the following steps:
s1, acquiring a current spectrogram of an instrument motor to identify fault symptoms of the instrument motor.
The method comprises the steps of obtaining a current spectrogram of an instrument motor after the instrument motor to be detected is electrified, and identifying fault signs of the instrument motor by analyzing image characteristic changes of the current spectrogram.
Referring to fig. 2, specifically, s101, based on a current frequency component image of a stator winding obtained by detecting an asymmetric motor rotor in an instrument, comparing the current frequency component image with a preset variable frequency band image to determine a rotor winding fault sign, wherein the preset variable frequency band is a variable frequency band with a size of 2sf and appears on both sides of a fundamental wave.
In the event of a fault on the rotor winding, the current frequency characteristic of the rotor winding can be expressed by the expression:
f s ＝[v(1-S)±S]f (v =1,3,5, …), where f s Representing the characteristic frequency of the current when the rotor winding fault occurs; s represents slip.
When a fault occurs on the rotor winding, the current of the stator winding has frequency components; at this point, it is just sensed for the motor rotor with asymmetry. And the symmetry of the rotor may even directly determine its size. In the spectrogram of the current signal, 2 sf-varying frequency bands appearing to have sizes appear on both sides of the fundamental wave.
S102, comparing a current harmonic image obtained by detecting a relatively static stator in the instrument with a preset current harmonic image to determine the air gap eccentricity fault sign, wherein the preset current harmonic is a fundamental wave and variable frequency bands appear on two sides of a main tooth harmonic.
After an eccentric fault occurs in an air gap of the motor, the generated characteristic parameter is abnormally changed into a magnetic flux waveform of the air gap, which is influenced by space and time. At this time, since the stator is relatively stationary, it can accurately sense the change of the current harmonics.
The formula can be used: f. of s ＝[1±(1-S)/p]f＝f±f r ；
In the formula f s Representing the characteristic frequency of the current when the air gap eccentric fault occurs; f. of r Representing the rotor rotational frequency; f represents a fundamental wave; p represents the number of pole pairs.
f s ＝[(R±1)(1-S)/p±1]f＝(R±1)f r ±f；
f s ＝[R(1-S)/p±1]f＝Rf r F, ± f; in the formula Rf r F denotes the main tooth harmonic.
In general, in the spectrogram of the current signal of the motor in the normal state, the existence of fundamental waves and main tooth harmonics can be seen. When the air gap eccentric fault occurs on the motor, in a spectrogram of a current signal, it is obviously found that frequency conversion bands appear on two sides of a fundamental wave and a main tooth harmonic wave, and the sizes of the frequency conversion bands are the same as the rotating frequency according to estimation.
S103, obtaining a test harmonic image based on the stator current and the air gap magnetic field detection, and comparing the test harmonic image with a preset harmonic image to determine the stator winding fault symptom, wherein the preset harmonic image is the nth harmonic and the mth harmonic of the current.
When a stator winding on the motor fails, the possible abnormal changes of the characteristic parameters are as follows: stronger harmonics are generated in both the stator current and the air gap magnetic field, and in the spectrogram of the current signal, the 3 rd harmonic and the 5 th harmonic of the preset harmonic image current are obviously found to have remarkable signs of increasing.
S2, obtaining the amplitude corresponding to the fault characteristic frequency component of the fault symptom and the amplitude of the fundamental component, and taking the ratio of the amplitude corresponding to the fault characteristic frequency component and the amplitude of the fundamental component as an input layer neuron of the BP neural network.
Referring to fig. 3, specifically, step S2 includes the following sub-steps:
s201, obtaining amplitudes corresponding to frequency components f, (1 + 2S) f, (1-2S) f, [1+ (1-S)/p ] f, [1- (1-S)/p ] f,3f,5f and amplitudes of fundamental wave components in a current spectrum diagram of an instrument motor;
s202, calculating a ratio between an amplitude corresponding to the fault characteristic frequency component and an amplitude of the fundamental component based on a ratio algorithm;
and S203, taking the ratio as an input layer neuron of the BP neural network.
And S3, acquiring a fault mode of the motor of the instrument to serve as an output layer neuron of the BP neural network, wherein the fault mode is used for representing the fault type of the motor of the instrument.
Referring to fig. 4, step S3 specifically includes the following sub-steps:
s301, acquiring various fault modes of an instrument motor;
s302, establishing a fault matrix based on the fault mode, wherein elements of a row vector of the fault matrix correspond to the fault mode;
s303, establishing an incidence relation between the fault matrix and an output layer neuron of the BP neural network, wherein [1,0,0], [0,1,0], [0,0,1] is used as a fault mode of the corresponding output layer neuron, zero-value elements represent no fault, and non-zero-value elements represent the corresponding fault mode.
Specifically, the input layer neuron is an amplitude value X = [ X ] corresponding to each fault characteristic frequency component 1 ,X2,…,X n ]And the ratio of the amplitudes of the fundamental wave components, the output layer neuron is various fault modes of the motor Y = [ Y ] 1 ,Y 2 ,…,Y m ] [10] 。
The failure modes corresponding to the output layer neurons may be, in order: (1) winding faults (short circuit, broken bars) occurring in the rotor; (2) eccentric failure of air gap; (3) Winding faults (short circuit, grounding) occur in the stator windings.
And S4, acquiring a standard training sample set and a test sample set, and obtaining a standard training sample test result corresponding to the standard training sample set and a test sample test result corresponding to the test sample set based on a BP neural network.
Referring to fig. 5, specifically, step S4 includes the following sub-steps:
s401, a standard training sample set and a test sample set are obtained. The standard training sample set comprises rotor winding fault samples, air gap eccentricity fault samples, stator winding fault samples and normal standard samples. The set of test samples is a set of samples from actual measurements.
S402, obtaining a standard training sample test result of the standard training sample set based on a BP neural network;
s403, obtaining a test sample test result of the standard training sample set based on the BP neural network;
s404, comparing the test result of the standard training sample with the test result of the test sample to obtain a comparison result;
s405, if all the comparison results are zero values, indicating that the instrument with the motor is normal and has no fault; if the comparison result has a non-zero value, the motor instrument is indicated to have a fault.
And S5, comparing the test result of the training sample with the test sample result to obtain the fault mode of the motor of the instrument.
The following illustrates the way in which the standard training sample set and the test sample set are established.
Mixing X 1 ,X 2 ,X 3 ,X 4 As a set of standard training samples.
Wherein, rotor winding fault sample: x 1 ＝[1,0.01,0.01,0.009,0.0004,0.003,0.05]；
Air gap eccentricity fault samples: x 2 ＝[1,0.0018,0.0018,0.049,0.041,0.023,0.027]；
Stator winding fault samples: x 3 ＝[1,0.01,0.01,0.134,0.006,0.09,0.027]；
Normal standard sample: x 4 ＝[1,0.01,0.01,0.004,0.004,0.035,0.03]。
The standard output corresponding to the sample data is the rotor winding fault standard output which is Y 1 ＝[1,0,0]；
Air gap eccentricity fault standard output is Y 2 ＝[0,1,0]；
The standard output of stator winding fault is Y 3 ＝[0,0,1]；
Normal no fault standard output is Y 4 ＝[0,0,0]。
Several groups of samples X obtained by actual measurement 5 ,X 6 ,X 7 ,X 8 As a test sample set:
X 5 ＝[0.89 0.00089 0.00089 0.00356 0.00356 0.03115 0.0267]；
X 6 ＝[0.9 0.009 0.009 0.0081 0.00036 0.027 0.045]；
X 7 ＝[0.9 0.00162 0.00162 0.0441 0.0369 0.0207 0.0243]；
X 8 ＝[0.9 0.0009 0.0009 0.01206 0.0054 0.081 0.0243]
the standard output corresponding to the test sample is Y 5 ＝[0,0,0]Indicating normal without failure;
Y 6 ＝[1,0,0]indicating a rotor winding fault;
Y 7 ＝[0,1,0]air gap eccentricity fault is indicated;
Y 8 ＝[0,0,1]indicating a stator winding fault.
The fault mode is associated with the fault matrix, and [1,0,0], [0,1,0], [0,0,1] is used as the fault mode of the corresponding neuron of the output layer, zero-value elements represent no fault, and non-zero-value elements represent the corresponding fault mode, so that the fault mode can be conveniently judged by workers and a computer.
Specifically, after determining the input layer neurons and the output layer neurons of the BP neural network, the method further includes:
and determining the number of nodes of the input layer and the output layer based on the required dimensions of the input vector and the output vector of the BP neural network.
The number of neurons in the hidden layer is determined based on a correlation empirical algorithm. The relevant empirical algorithm comprises
Wherein k is the number of samples, M is the number of neurons of a hidden layer of the BP neural network, and N is the number of neurons of an input layer;
and for all i > M, C is true M ＝0；
Wherein m is the number of neurons in the input layer, n is the number of neurons in the output layer, and a is [0,10 ]]Any constant within the interval.
The performance of the whole BP neural network can be basically determined by the number of the neurons of the hidden layer, generally, the more the number of the neurons of the hidden layer is, the better the network performance is, however, if the number of the neurons of the hidden layer is too much, the training time of the whole BP neural network is too long, the number of the neurons is determined through the formula, and the performance of the BP neural network is convenient to improve.
Transfer functions between layers of the BP neural network are determined. Specifically, a sigmoid function is adopted as the hidden layer transfer function. And taking a Log-sigmoid function as an output layer transfer function based on the output state of the neuron of the output layer. Generally, in the BP neural network, a function tansig with a value range of [ -1,1] and a function logsig with a value range of [0,1] are used as transfer functions, and the formula is:
the functional diagrams of the above equations refer to fig. 6 and 7.
In general, sigmoid functions are used as the hidden layer transfer functions, while the output layer uses linear functions as its transfer functions. If the output layer also uses the Sigmoid function as its transfer function, the output value is limited to (0,1) or (-1,1).
The embodiment of the application also discloses a readable storage medium which is stored with the readable storage medium and can be loaded by a processor to execute the hemodialysis equipment fault diagnosis method based on the BP neural network. Through the description of the foregoing embodiments, it is clear to those skilled in the art that the method of the foregoing embodiments may be implemented by software plus a necessary general hardware platform, and certainly may also be implemented by hardware, but in many cases, the former is a better implementation. Based on such understanding, the technical solutions of the present application or portions contributing to the prior art may be embodied in the form of a software product, where the computer software product is stored in a storage medium (such as a ROM/RAM, a magnetic disk, and an optical disk) as above, and includes several instructions to enable a device (which may be a mobile phone, a computer, a server, a controlled terminal, or a network device) to execute the method of each embodiment of the present application.
The above embodiments are preferred embodiments of the present application, and the protection scope of the present application is not limited by the above embodiments, so: equivalent changes in structure, shape and principle of the present application shall be covered by the protection scope of the present application.
Claims (10)
1. The hemodialysis equipment fault diagnosis method based on the BP neural network is characterized by comprising the following steps of:
acquiring a current spectrogram of an instrument motor to identify fault symptoms of the instrument motor;
obtaining an amplitude corresponding to a fault characteristic frequency component of a fault symptom and an amplitude of a fundamental component, and taking a ratio of the amplitude corresponding to the fault characteristic frequency component and the amplitude of the fundamental component as an input layer neuron of the BP neural network;
acquiring a fault mode of an instrument motor to serve as an output layer neuron of a BP neural network, wherein the fault mode is used for representing the fault type of the instrument motor;
acquiring a standard training sample set and a test sample set, and obtaining a standard training sample test result corresponding to the standard training sample set and a test sample test result corresponding to the test sample set based on a BP neural network;
and comparing the test result of the training sample with the test sample result to obtain the fault mode of the motor of the instrument.
2. The BP neural network-based hemodialysis device fault diagnosis method of claim 1, wherein the step of obtaining a current spectrogram of an instrument motor to identify signs of a fault of the instrument motor comprises:
comparing a current frequency component image of a stator winding obtained by detecting an asymmetric motor rotor in an instrument with a preset variable frequency band image to determine a rotor winding fault symptom, wherein the preset variable frequency band is a variable frequency band with the size of 2sf and appears on two sides of a fundamental wave;
comparing a current harmonic image obtained by detecting a relatively static stator in an instrument with a preset current harmonic image to determine an air gap eccentricity fault symptom, wherein the preset current harmonic is a fundamental wave and variable frequency bands appear on two sides of a main tooth harmonic;
and obtaining a test harmonic image based on the stator current and the air gap magnetic field detection, and comparing the test harmonic image with a preset harmonic image to determine the stator winding fault symptom, wherein the preset harmonic image is the nth harmonic and the mth harmonic increase of the current.
3. The BP neural network-based hemodialysis apparatus fault diagnosis method according to claim 1, wherein the step of obtaining the amplitude corresponding to the fault characteristic frequency component of the fault sign and the amplitude of the fundamental component, and using a ratio between the amplitude corresponding to the fault characteristic frequency component and the amplitude of the fundamental component as an input layer neuron of the BP neural network comprises:
acquiring amplitudes corresponding to frequency components f, (1 + 2S) f, (1-2S) f, [1+ (1-S)/p ] f, [1- (1-S)/p ] f,3f and 5f and amplitudes of fundamental wave components in a current spectrogram of an instrument motor;
calculating the ratio between the amplitude corresponding to the fault characteristic frequency component and the amplitude of the fundamental component based on a ratio algorithm;
and taking the ratio as an input layer neuron of the BP neural network.
4. The BP neural network-based hemodialysis apparatus fault diagnosis method according to claim 1, wherein the step of acquiring a fault mode of an instrument motor as an output layer neuron of the BP neural network comprises:
acquiring various fault modes of an instrument motor;
establishing a fault matrix based on the fault pattern, wherein elements of a row vector of the fault matrix correspond to the fault pattern; establishing the incidence relation between the fault matrix and the output layer neuron of the BP neural network, wherein
[1,0,0], [0,1,0], [0,0,1] as the failure mode for the corresponding output layer neurons, where zero value elements represent no failure and non-zero value elements represent the corresponding failure mode.
5. The BP neural network-based hemodialysis device fault diagnosis method according to claim 1, further comprising, after determining input layer neurons and output layer neurons of the BP neural network:
determining the number of nodes of an input layer and an output layer based on the dimension required by the input vector and the output vector of the BP neural network;
determining the number of neurons of the hidden layer based on a relevant empirical algorithm;
transfer functions between layers of the BP neural network are determined.
6. The BP neural network-based hemodialysis device fault diagnosis method of claim 5, wherein the relevant empirical algorithm comprises
and for all i > M, there is C M ＝0；
Wherein m is the number of neurons in the input layer, n is the number of neurons in the output layer, and a is [0,10 ]]Any constant within the interval.
7. The BP neural network-based hemodialysis device fault diagnosis method of claim 5, wherein the step of determining a transfer function between layers of the BP neural network comprises:
adopting a sigmoid function as a transfer function of a hidden layer;
and taking a Log-sigmoid function as an output layer transfer function based on the output state of the output layer neuron.
8. The BP neural network-based hemodialysis device fault diagnosis method according to claim 1, wherein the step of obtaining a standard training sample set and a test sample set, and obtaining a standard training sample test result corresponding to the standard training sample set and a test sample test result corresponding to the test sample set based on a BP neural network comprises:
acquiring a standard training sample set and a test sample set;
obtaining a standard training sample test result of the standard training sample set based on a BP neural network;
obtaining a test sample test result of the standard training sample set based on a BP neural network;
comparing the test result of the standard training sample with the test result of the test sample to obtain a comparison result;
if all the comparison results are zero values, indicating that the instrument with the motor is normal and has no fault;
if a non-zero value exists in the comparison result, the motor instrument has a fault.
9. The BP neural network-based hemodialysis device fault diagnosis method of claim 8, wherein the standard training sample set comprises rotor winding fault samples, air gap eccentricity fault samples, stator winding fault samples and normal standard samples;
the set of test samples is a set of samples from actual measurements.
10. A computer readable storage medium storing at least one instruction, at least one program, a set of codes, or a set of instructions, which is loaded and executed by a processor to implement the BP neural network based hemodialysis device fault diagnosis method according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210416019.9A CN115327373B (en) | 2022-04-20 | 2022-04-20 | Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210416019.9A CN115327373B (en) | 2022-04-20 | 2022-04-20 | Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115327373A true CN115327373A (en) | 2022-11-11 |
| CN115327373B CN115327373B (en) | 2023-04-18 |
Family
ID=83915819
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210416019.9A Active CN115327373B (en) | 2022-04-20 | 2022-04-20 | Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115327373B (en) |
Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102721545A (en) * | 2012-05-25 | 2012-10-10 | 北京交通大学 | Rolling bearing failure diagnostic method based on multi-characteristic parameter |
| CN104166095A (en) * | 2014-08-29 | 2014-11-26 | 东南大学 | Fault information fusion diagnosis method based on double-edge linear motor |
| CN105242173A (en) * | 2015-09-14 | 2016-01-13 | 吉林大学 | Frequency domain electromagnetic method-based grounding grid fault automatic diagnosis method |
| CN105487009A (en) * | 2015-11-19 | 2016-04-13 | 上海电机学院 | Motor fault diagnosis method based on k-means RBF neural network algorithm |
| US20160282416A1 (en) * | 2015-03-26 | 2016-09-29 | The University Of Akron | System and method for iterative condition monitoring and fault diagnosis of electric machines |
| KR101797402B1 (en) * | 2016-12-07 | 2017-11-15 | 울산대학교 산학협력단 | Method and apparatus for diagnosing machine defects |
| CN108768550A (en) * | 2018-06-21 | 2018-11-06 | 中国人民解放军国防科技大学 | Wide-band transmitter nonlinear modeling method based on dynamic multi-core bandwidth generalized regression neural network algorithm |
| CN109798996A (en) * | 2019-01-21 | 2019-05-24 | 岱特智能科技（上海）有限公司 | A kind of system for detecting temperature |
| US20190242936A1 (en) * | 2018-02-05 | 2019-08-08 | Wuhan University | Fault diagnosis method for series hybrid electric vehicle ac/dc converter |
| CN110223195A (en) * | 2019-05-22 | 2019-09-10 | 上海交通大学 | Distribution network failure detection method based on convolutional neural networks |
| CN110766143A (en) * | 2019-10-31 | 2020-02-07 | 上海埃威航空电子有限公司 | Equipment fault intelligent diagnosis method based on artificial neural network |
| CN112003272A (en) * | 2020-08-11 | 2020-11-27 | 东北电力大学 | Power system frequency safety control method based on convolutional neural network |
| CN113237619A (en) * | 2021-04-16 | 2021-08-10 | 江苏方天电力技术有限公司 | Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration |
| CN113568396A (en) * | 2021-08-04 | 2021-10-29 | 江苏矿星智能科技有限公司 | Mine main ventilator fault diagnosis method based on neural network |
| US11220999B1 (en) * | 2020-09-02 | 2022-01-11 | Palo Alto Research Center Incorporated | Deep hybrid convolutional neural network for fault diagnosis of wind turbine gearboxes |
| CN114118383A (en) * | 2021-12-09 | 2022-03-01 | 之江实验室 | Multi-synaptic plasticity pulse neural network-based fast memory coding method and device |
| CN114154400A (en) * | 2021-11-15 | 2022-03-08 | 中国人民解放军63963部队 | Unmanned vehicle health state detection system and detection method |
| US20220096003A1 (en) * | 2015-07-17 | 2022-03-31 | Chao-Lun Mai | Method, apparatus, and system for automatic and adaptive wireless monitoring and tracking |
| WO2022067562A1 (en) * | 2020-09-29 | 2022-04-07 | 西门子股份公司 | Method and device for diagnosing fault arc, and computer-readable storage medium |
-
2022
- 2022-04-20 CN CN202210416019.9A patent/CN115327373B/en active Active
Patent Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102721545A (en) * | 2012-05-25 | 2012-10-10 | 北京交通大学 | Rolling bearing failure diagnostic method based on multi-characteristic parameter |
| CN104166095A (en) * | 2014-08-29 | 2014-11-26 | 东南大学 | Fault information fusion diagnosis method based on double-edge linear motor |
| US20160282416A1 (en) * | 2015-03-26 | 2016-09-29 | The University Of Akron | System and method for iterative condition monitoring and fault diagnosis of electric machines |
| US20220096003A1 (en) * | 2015-07-17 | 2022-03-31 | Chao-Lun Mai | Method, apparatus, and system for automatic and adaptive wireless monitoring and tracking |
| CN105242173A (en) * | 2015-09-14 | 2016-01-13 | 吉林大学 | Frequency domain electromagnetic method-based grounding grid fault automatic diagnosis method |
| CN105487009A (en) * | 2015-11-19 | 2016-04-13 | 上海电机学院 | Motor fault diagnosis method based on k-means RBF neural network algorithm |
| KR101797402B1 (en) * | 2016-12-07 | 2017-11-15 | 울산대학교 산학협력단 | Method and apparatus for diagnosing machine defects |
| US20190242936A1 (en) * | 2018-02-05 | 2019-08-08 | Wuhan University | Fault diagnosis method for series hybrid electric vehicle ac/dc converter |
| CN108768550A (en) * | 2018-06-21 | 2018-11-06 | 中国人民解放军国防科技大学 | Wide-band transmitter nonlinear modeling method based on dynamic multi-core bandwidth generalized regression neural network algorithm |
| CN109798996A (en) * | 2019-01-21 | 2019-05-24 | 岱特智能科技（上海）有限公司 | A kind of system for detecting temperature |
| CN110223195A (en) * | 2019-05-22 | 2019-09-10 | 上海交通大学 | Distribution network failure detection method based on convolutional neural networks |
| CN110766143A (en) * | 2019-10-31 | 2020-02-07 | 上海埃威航空电子有限公司 | Equipment fault intelligent diagnosis method based on artificial neural network |
| CN112003272A (en) * | 2020-08-11 | 2020-11-27 | 东北电力大学 | Power system frequency safety control method based on convolutional neural network |
| US11220999B1 (en) * | 2020-09-02 | 2022-01-11 | Palo Alto Research Center Incorporated | Deep hybrid convolutional neural network for fault diagnosis of wind turbine gearboxes |
| WO2022067562A1 (en) * | 2020-09-29 | 2022-04-07 | 西门子股份公司 | Method and device for diagnosing fault arc, and computer-readable storage medium |
| CN113237619A (en) * | 2021-04-16 | 2021-08-10 | 江苏方天电力技术有限公司 | Fault early warning method, device, equipment and storage medium for variable-speed rotating machinery vibration |
| CN113568396A (en) * | 2021-08-04 | 2021-10-29 | 江苏矿星智能科技有限公司 | Mine main ventilator fault diagnosis method based on neural network |
| CN114154400A (en) * | 2021-11-15 | 2022-03-08 | 中国人民解放军63963部队 | Unmanned vehicle health state detection system and detection method |
| CN114118383A (en) * | 2021-12-09 | 2022-03-01 | 之江实验室 | Multi-synaptic plasticity pulse neural network-based fast memory coding method and device |
Non-Patent Citations (2)
| Title |
|---|
| OLIVIER JANSSENS 等: "Convolutional Neural Network Based Fault Detection for Rotating Machinery", 《JOURNAL OF SOUND AND VIBRATION》 * |
| 郭俊: "变转速工况下直驱式永磁同步风力发电机轴承故障诊断方法研究", 《中国博士学位论文全文数据库 工程科技Ⅱ辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115327373B (en) | 2023-04-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bessous et al. | Diagnosis of bearing defects in induction motors using discrete wavelet transform | |
| JP6223461B2 (en) | Diagnosis method of electromechanical system based on impedance analysis | |
| Amanuel et al. | Comparative analysis of signal processing techniques for fault detection in three phase induction motor | |
| CN116433009A (en) | Abnormality monitoring method and device for power transformation equipment and storage medium | |
| US10976379B2 (en) | Technique to detect motor leakage flux anomalies | |
| WO2018066267A1 (en) | Rotating machine abnormality detection device and method, and rotating machine | |
| JP2024019278A (en) | Abnormality diagnosis device and abnormality diagnosis method | |
| Jokic et al. | The analysis of vibration measurement and current signature in motor drive faults detection | |
| JP2020148461A (en) | Signal processing device and signal processing method | |
| CN115327373B (en) | Hemodialysis equipment fault diagnosis method based on BP neural network and storage medium | |
| JPH0843257A (en) | Method for diagnosing planetary gear | |
| CN110161344A (en) | Transformer vibration sound detection signal filtering method and system based on Scale Decomposition | |
| Pietrzak et al. | Application of spectral and wavelet analysis of stator current to detect angular misalignment in pmsm drive systems | |
| KR20210120049A (en) | Anomaly diagnosis device and abnormality diagnosis method | |
| CN112671307B (en) | Surface-mounted permanent magnet synchronous motor demagnetization degree evaluation method and device | |
| JPH04204021A (en) | Apparatus for diagnosing vibration and sound of rotating machine | |
| JP7381390B2 (en) | Abnormality diagnosis device and maintenance management system | |
| Arashloo et al. | A novel broken rotor bar fault detection method using park's transform and wavelet decomposition | |
| Babu et al. | Review on various signal processing techniques for predictive maintenance | |
| Bessous et al. | Application of new quantitative and qualitative study based on DWT method to diagnose the eccentricity fault in induction motors | |
| EP4194979A1 (en) | System and method for monitoring condition of an electromechanical system | |
| WO2022190795A1 (en) | Abnormality detection device and method | |
| CN111579990B (en) | Induction motor rotor broken bar fault detection method based on stator current absolute phase | |
| JPH0731076B2 (en) | Abnormality diagnosis method for rotating machinery | |
| WO2022107100A1 (en) | Method and system for auto-detecting induction motor fault |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |